The blood-brain barrier (BBB), while providing effective protection to the brain against circulating toxins, also creates major difficulties in the pharmacological treatment of brain diseases such as Alzheimer's disease, Parkinson's disease, and brain cancer. Most charged molecules, and most molecules over 700 Daltons in size, are unable to pass through the barrier, and smaller molecules may be conjugated in the liver. These factors create major difficulties in the pharmacological treatment of diseases of the brain and central nervous system (CNS), such as Alzheimer's disease, Parkinson's disease, bacterial and viral infections and cancer.
Many therapeutic agents for the treatment of diseases and disorders of the brain and CNS are sufficiently hydrophilic to preclude direct transport across the BBB. Furthermore, these drugs and agents are susceptible to degradation in the blood and peripheral tissues that increase the dose necessary to achieve a therapeutically effective serum concentration. The present invention provides a method of delivering therapeutic agents to the BBB by encapsulating the therapeutic agent in an artificial low-density lipoprotein particle (LDL). The LDL of this invention facilitates transport of therapeutic agents across the BBB by transcytosis. Since most drugs and therapeutic agents that are too hydrophilic to cross the BBB are also too hydrophilic to be incorporated into an LDL particle, the present invention provides a method for producing conjugates of the therapeutic agents with an LDL component that facilitates its incorporation into an LDL particles, transport across the BBB and subsequent release of the therapeutic agent into the cell.
Prior methods for delivery drugs across the BBB involve three general categories: (1) liposome-based methods, where the therapeutic agent is encapsulated within the carrier; (2) synthetic polymer-based methods, where particles are created using synthetic polymers to achieve precisely-defined size characteristics; and (3) direct conjugation of a carrier to a drug, where the therapeutic agent is covalently bound to a carrier such as insulin.
A. Liposomes
Liposomes are small particles that form spontaneously when phospholipids are sonicated in aqueous solution, and consist of a symmetrical lipid bilayer configured as a hollow sphere surrounding an aqueous environment. This has appeal as a means of transporting water-soluble drugs through the cell membrane, as the phospholipid can be absorbed in the plasma membrane, which automatically releases the contents of the liposomes into the cytosol. More successful variations of this technique include the use of cationic lipids, which can cooperatively create nanopores in the membrane. Cationic lipids are extensively used in cell culture to introduce water-soluble materials such as DNA molecules into cultured cells for experimentation.
Liposomes are attractive for transporting drugs across the BBB because of their large carrying capacity. However, liposomes are generally too large to effectively cross the BBB, are inherently unstable, and their constituent lipids are gradually lost by absorption by lipid-binding proteins in the plasma. For example, in some studies, the large size of the liposomes used produced microembolisms that gave a false impression of brain uptake. In some studies, liposomes were co-injected with Polysorbate 80, a detergent that can disrupt the BBB, as a stabilizing agent. The disruption of the BBB by the Polysorbate 80 in these studies may be responsible for any observed transport across the BBB.
Consequently, liposomes have had a checkered history as vehicles for transporting drugs across the BBB. Several attempts have been made to direct the liposome to particular cellular targets. Peptidomimetic mAbs that target endogenous receptors of the BBB have been used to target pegylated immunoliposomes to various BBB receptors, with the aim of achieving receptor-mediated uptake. However, this approach also requires the expensive production, testing and governmental approval of monoclonal antibodies. Because mAbs are typically produced in mice, and are susceptible to degradation, introduction of a peptidomimetic mAb would not only face significant regulatory obstacles but would prove difficult to deploy in a patient environment.
Immunoliposomes, for example, have been constructed in a process that involves covalent attachment of monoclonal antibodies (mAbs) to the surface of the liposome. Since these immunoliposomes are immediately coated with plasma proteins that trigger uptake by the reticuloendothelial systems (RES), a system that avidly destroys mAb-conjugated liposomes, immunoliposomes have been treated with polyethylene glycol (PEG) in a process known as pegylation. Unfortunately, the PEG molecules interfere with the mAb, rendering them non-specific due to steric interference. Huwyler et al. (1996) Proc. Nat'l. Acad. Sci USA 93: 14164-14169 avoided this problem by creating immunoliposomes with a maleimide moiety at the tip of the PEG tail, which could be conjugated with a thiolated mAb. These pegylated OX26 immunoliposomes, which were prepared with daunomycin in their interiors, were shown to be more stable in plasma than the free therapeutic agent or plain, unpegylated liposomes. Confocal microscopy, however, has shown that although the liposomes were endocytosed into rat brain capillaries, they did not reach brain cells and remained attached to endothelial cells. Thus, pegylated and maleimide-treated liposomes appear to be relatively ineffective as drug delivery vehicles.
In 1997, Dehouck et al. discovered that the LDL receptors, which binds ApoE, is involved in transcytosis of LDL across the BBB. In a series of three publications, Versluis et al. described the use of ApoE-enriched liposomes to deliver daunorubicin to cancer cells in mice. ApoE was selected as an LDL-receptor targeting protein based on the finding that tumor cells express high levels of LDL receptors on their membranes. Versluis et al. (1998) also proposed using natural LDL, but this experiment was not attempted and subsequent papers focused exclusively on ApoE-enriched liposomes. Versluis et al. (1999) examined the tissue distribution of daunorubicin, but there are no data related to brain uptake, indicating that this method was not envisaged as a means for transporting daunorubicin across the BBB.
Additionally, the conjugation chemistry used by Versluis et al. is different from that used in the present Invention. To anchor the drug to the liposome membrane, the authors coupled 3α-O-(oleoyl)-4β-cholanic acid (an ester of lithocholic acid) to the tetrapeptide Ala-Leu-Ala-Leu, which was in turn covalently linked to the hydrophilic anti-tumor agent daunorubicin. Thus, tumors were treated with conjugated, not free, daunorubicin. Although lithocholic acid is a steroid that already contains an activatable acid group, the acid group is located on the steroid side chain instead of the 3-OH position, which results in a reaction product with less desirable features. Free daunorubicin can be produced only after cleavage by proteases fund in the highly acidic lysosome, which would expose the conjugated during or agent to degradation by proteases, acid and other hydrolytic enzymes. The therapeutic agent would then be released into the intralysosomal space where it could undergo further degradation and expulsion from the cell.
In contrast, the conjugates of the present invention preferably provides for attachment of the therapeutic agent via an ester linkage that can be easily cleaved in the cytosol and consequently escape the harsh lysosomal conditions needed by the method of Versluis et al. Thus, a therapeutic agent conjugated by the present method would be more likely to survive the journey to its target and to be released at the target in an efficient manner. It is also more likely to be transported across the BBB than a liposome.
The method of Versluis et al. also requires a large number of solid-phase peptide chemistry steps to synthesize the tetrapeptide, and several additional steps to conjugate it with FMOC and react the conjugate with lithocholic acid and finally with the drug. The present invention uses a much smaller number of steps, each of which produces nearly quantitative yield. Thus the present invention also offers improved efficiency and lower cost.
Other liposomal formulations of doxyrubicin are currently in clinical use as possible treatments for cancer; however, no products have been introduced that use LDL.
Demeule et al. found that the protein melanotransferrin (p97) is transported by transcytosis across the BBB and concluded that an LDL receptor was involved, suggesting that this protein be employed as a drug delivery system.
B. Synthetic Polymers
Synthetic polymers such as poly(butyl cyanoacrylate) or polyacrylamide covered with Polysorbate 80 have also been tried. These polymers are appealing because the particles are sufficiently hydrophilic to be water-soluble, yet are able to maintain their structural form for long periods, which protects the therapeutic agent from uptake into the liver and kidney where it is subject to natural detoxification process. In both cases, uptake is generally supposed to occur by passive diffusion across the cell membrane or as a defensive uptake by clathrin-coated vesicles. In the former case, the therapeutic agent is then trapped in an endothelial cell, where it is not much closer to the target than before, whereas in the latter case, the therapeutic agent is transported to a lysosome, which is a highly acidic compartment in the cell containing proteases and other digestive enzymes analogous to stomach contents. Thus, in the latter case, the therapeutic agent must remain stable throughout more extreme conditions. In neither case is the drug carried across the cell and ejected into the brain parenchyma, which is the desired result. Thus, it is not surprising that neither of these two methods has achieved much clinical use.
Numerous researchers have tried various modifications of the approaches described above to improve carrier uptake across the BBB with limited success. For example, Kreuter et al. (2002) J. Drug Target 10(4): 317-25 engineered synthetic particles that contained various apolipoproteins that would bind to the apolipoprotein receptors located at the BBB. They demonstrated transport of drugs bound to poly(butyl cyanoacrylate) nanoparticles and coated them with Polysorbate 80. Uptake required coating with Polysorbate 80, ApoE or ApoB. Apolipoproteins All, Cll, or J coatings did not work. However, because these nanoparticles are not naturally occurring, they may have undesirable side effects. Acrylate polymers are particularly notorious for initiating autoimmune responses; the chemically-related polymer poly(acrylamide) is often used as an adjuvant.
Alyaudin et al. (2001) J. Drug Target 9(3): 209-21 used poly(butylcyanoacrylate) nanoparticles overcoated with Polysorbate 80 to transport [3H]-dalargin across the BBB and surmised the process was one of endocytosis followed by possible transcytosis. This polymer may have immunological complications as well.
C. Therapeutic Agent Conjugates
Direct conjugation of pharmacological agents with the substances that can be transported across the BBB, such as insulin, has also been attempted. Insulin and insulin-like growth factors are known to cross the blood brain barrier by specialized facilitated diffusion systems. (Reinhardt et al. (1994) Endocrinology 135(5): 1753-1761). Insulin is taken up by transcytosis mediated by the endothelial insulin receptor (Pardridge et al. (1986) Ann. Intern. Med. 105(1): 82-95). Specific transporters also exist for glucose and for large amino acids such as tryptophan. However, the specificity of the insulin transporter has proved to be too high to allow pharmacological agents covalently linked to insulin to cross into the brain. Similar results have been obtained with glucose and amino acid conjugates, whose uptake has been observed to obey the same general principles as other low-molecular weight substances, with only uncharged molecules below 700 Da achieving significant access to the brain. The inconvenience in devising chemical syntheses of conjugated forms of biomolecules, the risk of creating unanticipated toxic effects, and the likely necessity of obtaining FDA approval for an entirely novel compound has dampened enthusiasm for this approach.
Transport vectors, which are proteins such as cationized albumin, or the OX26 monoclonal antibody to the transferrin receptor undergo absorptive-mediated and receptor-mediated transcytosis through the BBB, respectively. These have been used to transport small amounts of drug. This process, suffers from the high expense and difficulty of producing monoclonal antibodies and cationized albumin and is not applicable to other types of molecules. Also, cationized proteins have been shown to be toxic due to their immunogenicity and the formation of immune complexes that are deposited in the kidney.
Wu et al. (2002) J. Drug Target 10(3): 239-45 showed transport of human basic fibroblast growth factor (bFGF), a protein neuroprotective agent, across the BBB using a drug delivery vector consisting of a conjugate of streptavidin (SA) and the murine OX26 monoclonal antibody against the rat transferrin receptor, and the conjugate of biotinylated bFGF (bio-bFGF) bound to a vector designated bio-bFGF/OX26-SA. Although they showed avid uptake of [125I] labeled bio-bFGF into peripheral organs, only 0.01% of the injected dose was taken up per gram brain. Also, this procedure requires covalent modification of the drug, and may be useful only for limited classes of drugs. The carrier also contains mouse monoclonal antibodies as a component, which would cause an immune response in the patient.
Kang et al. (2000) J. Drug Target 8(6): 425-34 also used an avidin-biotin linked chimeric peptide to transport a peptide across the BBB but achieved only 0.12% of the injected dose taken up per gram of tissue. Kang and Pardridge (Pharm. Res. 11: 1257-1264) conjugated cationized human serum albumin with neutral light avidin, and then bound it to radiolabeled biotin. The biotin/cHSA/NLA complex was stable in blood for up to 24 h, but the conjugate was selectively degraded in brain to release free biotin. As mentioned above, cationized proteins have been shown to be toxic due to their immunogenicity.
Cationized monoclonal antibodies (mAbs) have also been used. Pardridge (J. Neurochem. 70: 1781-2) showed by confocal microscopy that the native humanized 4D5 MAb crossed the BBB by absorptive-mediated transcytosis, but only after cationization of the protein. This process, however, suffers from the high expense of producing and chemically modifying monoclonal antibodies and is not applicable to other types of molecules.
Witt et al. (2000) J. Pharmacol. Exp. Ther. 295(3): 972-8 used insulin to transport delta-opioid receptor-selective peptide D-penicillamine (DPDPE), a Metenkephalin analog, across the BBB. Insulin, however, presents numerous hazards that limit its use as a therapeutic strategy. Also, other researchers have found the insulin receptor to be extremely selective. Thus, in addition to the difficulty in producing chimeric peptides, this strategy is limited to a narrow class of pharmaceutical agents.
Other researchers have attempted to conjugate drugs to glucose, for instance using glycopeptides. However, no significant transport of any glycopeptide via the BBB Glut1 transporter has ever been demonstrated. Attempts to use the high-transport rate of carrier-mediated transporters such as the Glut1 glucose transporter, the choline transporter, or the LAT1 large amino acid transporter have foundered on the problem that carrier transporters are too selective to accept conjugated substrates. They also suffer from the problem that p-glycoprotein, a member of the multidrug resistance gene, rapidly acts to actively remove many small molecules, including any drugs that manage to get across the barrier, from the brain.
In addition to the LDL receptor, the BBB also contains type II scavenger receptors (SR), which bind LDL with high affinity. The scavenger receptor is particularly good with modified forms of LDL such as acetylated LDL. Binding to the SR results only in endocytosis and not the desired transcytosis. Rigotti et al. (1995) J. Biol. Chem. 270: 16221-4 found that acetylated LDL is not transported across the BBB, whereas cationized bovine IgG was more effective Bickel et al. (1993) Adv. Drug. Del. Rev. 10: 205-245. The failure to demonstrate transcytosis with acetylated LDL discouraged many researchers from attempting further experiments with LDL.
Protter et al. (WO 87/02061) describe a drug delivery system that uses peptides derived from apolipoproteins, such as ApoE and ApoB, which are covalently attached to the pharmaceutical agent, or to a carrier containing the agent. However, the use of molecular conjugates would only be limited to a small number of drug classes, and subject to many of the same problems discussed above.
Müller et al. (U.S. Pat. No. 6,288,040) describe the use of synthetic poly(butyl cyanoacrylate) particles to which ApoE molecules are covalently bound. The surface of the particles are further modified by surfactants or covalent attachment of hydrophilic polymers. As stated above, because these particles are not naturally occurring, they may have a variety of undesirable side effects.
Samain et al. (WO 92/21330) describe the use of synthetic particulate carriers containing lipids that are covalently attached to a solid, hydrophilic core and that also contain ApoB for delivery of substances to tumors or macrophages. However, they do not disclose any utility of such vectors for delivering drugs across the BBB.